Remove UserInfo (#751)

src/Algorithm/branching/branchingalgo.jl

@@ -42,11 +42,8 @@ Chooses the best candidate according to a selection criterion and generates the
     selection_criterion::AbstractSelectionCriterion = MostFractionalCriterion()
     rules = [PrioritisedBranchingRule(SingleVarBranchingRule(), 1.0, 1.0)]
     int_tol = 1e-6
-    root_user_info::AbstractNodeUserInfo = DummyUserInfo()
 end
 
-get_root_user_info(algo::Branching) = algo.root_user_info
-
 get_selection_nb_candidates(::Branching) = 1
 
 struct BranchingContext{SelectionCriterion<:AbstractSelectionCriterion} <: AbstractDivideContext
@@ -114,11 +111,8 @@ restrict the number of retained candidates until only one is left.
     selection_criterion::AbstractSelectionCriterion = MostFractionalCriterion()
     verbose = true
     int_tol = 1e-6
-    root_user_info::AbstractNodeUserInfo = DummyUserInfo()
 end
 
-get_root_user_info(algo::StrongBranching) = algo.root_user_info
-
 ## Implementation of Algorithm API.
 
 # StrongBranching does not use any storage unit itself, 

src/Algorithm/conquer.jl

@@ -36,10 +36,11 @@ struct NodeFinalizer
 end
 
 ####################################################################
-#                      BeforeCutGenUserAlgo
+#                      BeforeCutGenAlgo
 ####################################################################
 
-struct BeforeCutGenUserAlgo <: AbstractConquerAlgorithm
+"Algorithm called before cut generation."
+struct BeforeCutGenAlgo <: AbstractConquerAlgorithm
     algorithm::AbstractOptimizationAlgorithm
     name::String
 end
@@ -103,7 +104,7 @@ Parameters :
 @with_kw struct ColCutGenConquer <: AbstractConquerAlgorithm 
     stages::Vector{ColumnGeneration} = [ColumnGeneration()]
     primal_heuristics::Vector{ParameterizedHeuristic} = [ParamRestrictedMasterHeuristic()]
-    before_cutgen_user_algorithm::Union{Nothing, BeforeCutGenUserAlgo} = nothing
+    before_cutgen_user_algorithm::Union{Nothing, BeforeCutGenAlgo} = nothing
     node_finalizer::Union{Nothing, NodeFinalizer} = nothing
     preprocess = nothing
     cutgen = CutCallbacks()
@@ -202,24 +203,23 @@ Returns `true` if the conquer algorithm continues.
 function run_colcutgen!(ctx::ColCutGenContext, env, reform, node_state)
     nb_cut_rounds = 0
     run_conquer = true
-    cuts_were_added = true
-    master_changed = true
+    master_changed = true # stores value returned by the algorithm called before cut gen.
+    cuts_were_added = true # stores value returned by cut gen.
     while run_conquer && (cuts_were_added || master_changed)
         for (stage_index, colgen) in Iterators.reverse(Iterators.enumerate(ctx.params.stages))
-            # print stage_index
+            # TODO: print stage_index
             run_conquer = run_colgen!(ctx, colgen, env, reform, node_state)
             if !run_conquer
                 return false
             end
         end
 
+        master_changed = false
         before_cutgen_user_algorithm = ctx.params.before_cutgen_user_algorithm
         if !isnothing(before_cutgen_user_algorithm)
             master_changed = run_before_cutgen_user_algo!(
                 ctx, before_cutgen_user_algorithm, env, reform, node_state
             )
-        else
-            master_changed = false
         end
 
         sol = get_best_lp_primal_sol(node_state)
@@ -392,12 +392,7 @@ end
 function run!(algo::ColCutGenConquer, env::Env, reform::Reformulation, input::AbstractConquerInput)
     !run_conquer(input) && return
     ctx = new_context(type_of_context(algo), algo, reform, input)
-    node = get_node(input)
-    parent = get_parent(node)
-    info = isnothing(parent) ? get_user_info(node) : get_user_info(parent)
-    MathProg.set_user_info!(reform, MathProg.copy_info(info))
     run_colcutgen_conquer!(ctx, env, reform, input)
-    set_user_info!(node, MathProg.get_user_info(reform))
     return
 end
 

src/Algorithm/interface.jl

@@ -111,8 +111,6 @@ abstract type AbstractDivideAlgorithm <: AbstractAlgorithm end
 # divide algorithms are always manager algorithms (they manage storing and restoring units)
 ismanager(algo::AbstractDivideAlgorithm) = true
 
-get_root_user_info(::AbstractDivideAlgorithm) = DummyUserInfo()
-
 @mustimplement "DivideAlgorithm" run!(::AbstractDivideAlgorithm, ::Env, ::AbstractModel, ::AbstractDivideInput) 
 
 # this function is needed to check whether the best primal solution should be copied to the node optimization state

src/Algorithm/treesearch/branch_and_bound.jl

@@ -9,7 +9,6 @@ mutable struct Node <: AbstractNode
     branchdescription::String
     records::Records
     conquerwasrun::Bool
-    user_info::AbstractNodeUserInfo
 end
 
 getdepth(n::Node) = n.depth
@@ -23,9 +22,6 @@ set_records!(n::Node, records) = n.records = records
 
 get_branch_description(n::Node) = n.branchdescription # printer
 
-set_user_info!(n::Node, info::AbstractNodeUserInfo) = n.user_info = info
-get_user_info(n::Node) = n.user_info
-
 # Priority of nodes depends on the explore strategy.
 get_priority(::AbstractExploreStrategy, ::Node) = error("todo")
 get_priority(::DepthFirstStrategy, n::Node) = -n.depth
@@ -35,7 +31,7 @@ get_priority(::BestDualBoundStrategy, n::Node) = get_ip_dual_bound(n.optstate)
 function Node(node::SbNode)
     return Node(
         node.depth, node.parent, node.optstate, node.branchdescription,
-        node.records, node.conquerwasrun, get_user_info(node.parent)
+        node.records, node.conquerwasrun
     )
 end
 
@@ -139,8 +135,7 @@ end
 function new_root(sp::BaBSearchSpace, input)
     nodestate = OptimizationState(getmaster(sp.reformulation), input, false, false)
     return Node(
-        0, nothing, nodestate, "", create_records(sp.reformulation), false,
-        get_root_user_info(sp.divide)
+        0, nothing, nodestate, "", create_records(sp.reformulation), false
     )
 end
 

src/MathProg/MathProg.jl

@@ -67,7 +67,7 @@ export Problem, set_initial_dual_bound!, set_initial_primal_bound!,
 # Methods related to Reformulation
 export Reformulation, getmaster, add_dw_pricing_sp!, add_benders_sep_sp!, get_dw_pricing_sps,
     set_reformulation!, get_benders_sep_sps, get_dw_pricing_sp_ub_constrid,
-    get_dw_pricing_sp_lb_constrid, setmaster!, get_user_info, set_user_info!
+    get_dw_pricing_sp_lb_constrid, setmaster!
 
 # Methods related to formulations
 export AbstractFormulation, Formulation, create_formulation!, getvar, getvars,

src/MathProg/reformulation.jl

@@ -39,9 +39,6 @@ ClB.getstorage(reform::Reformulation) = reform.storage
 
 # methods specific to Formulation
 
-set_user_info!(reform::Reformulation, info::AbstractNodeUserInfo) = reform.node_user_info = info
-get_user_info(reform::Reformulation) = reform.node_user_info
-
 """
     getobjsense(reformulation)
 

src/MathProg/types.jl

@@ -133,13 +133,3 @@ used by algorithms. A formulation contains one storage unit
 per storage unit type used by algorithms.    
 """
 abstract type AbstractFormulation <: AbstractModel end
-
-"Sets the user info stored at the formulation"
-function set_user_info!(::AbstractFormulation, ::AbstractNodeUserInfo)
-    return  # Fall back that does nothing
-end
-
-"Gets the user info stored at the formulation"
-function get_user_info(::AbstractFormulation)
-    return DummyUserInfo()  # Fall back that gets a dummy user info
-end

test/integration/improve_relax_callback.jl

@@ -48,7 +48,7 @@ function ClA.get_branching_candidate_units_usage(::ClA.SingleVarBranchingCandida
     return units_to_restore
 end
 
-ClA.ismanager(::ClA.BeforeCutGenUserAlgo) = false
+ClA.ismanager(::ClA.BeforeCutGenAlgo) = false
 ClA.ismanager(::ImproveRelaxationAlgo) = false
 
 # Don't need this because `ToyNodeInfo` is bits
@@ -72,7 +72,7 @@ function ClA.get_units_usage(algo::ImproveRelaxationAlgo, reform::ClMP.Reformula
     return units_usage
 end
 
-function ClA.get_child_algorithms(algo::ClA.BeforeCutGenUserAlgo, reform::ClMP.Reformulation)
+function ClA.get_child_algorithms(algo::ClA.BeforeCutGenAlgo, reform::ClMP.Reformulation)
     println("\e[42m THIS IS A CALL TO get_child_algorithm \e[00m")
     child_algos = Tuple{ClA.AbstractAlgorithm, ClMP.AbstractModel}[]
     push!(child_algos, (algo.algorithm, reform))
@@ -108,7 +108,7 @@ function test_improve_relaxation(; do_improve::Bool)
                                 ))
                                 ],
                     primal_heuristics = [],
-                    before_cutgen_user_algorithm = ClA.BeforeCutGenUserAlgo(
+                    before_cutgen_user_algorithm = ClA.BeforeCutGenAlgo(
                             ImproveRelaxationAlgo(
                                 userfunc = call_improve_relaxation
                             ), 
@@ -174,8 +174,7 @@ function test_improve_relaxation(; do_improve::Bool)
 
         # increment the user info value for testing
         if !do_improve
-            info = ClMP.get_user_info(reform)
-            ClMP.set_user_info!(reform, ToyNodeInfo(info.value + 111))
+            unit.value += 111
         end
         return
     end